Dynamic traffic routing using ant based control

This paper tackles the problem of dynamic routing for traffic in a city. We built a simulation environment with a routing system that guides the vehicles through the city using the fastest way, taking into account the load on the roads. A version of the AntNet algorithm is applied for routing the traffic. The artificial ants are moving in a virtual street network where each node corresponds to a crossroad. Each node has a routing table used for guiding the cars. This model is supplemented with actual data from the traffic by the cars themselves. This enables the agents to divert the traffic from congested routes and improves the traveling time. The simulation environment makes possible to see the effect in different cities or accident circumstances, especially in crisis situations, when, because of a disaster, multiple roads become unavailable or are heavily congested, and most of the drivers are disoriented.